Variables Used in the Probit Analysis

Variables used in the probit analysis are listed below. They consist of the dependent variable, information available, risk preference, farm size, farmer characteristics, and income.

Dependent Variable: The dependent binary variable, VSHX, indicates whether or not the breeder adopts VSH queens with the question, "Do you breed or sell queens?” A list of independent variables is described in Table 3.1. Included is risk preference, information available to the bee breeders such as being a member of a local beekeeping club, farm attributes such as number of colonies, income and demographics in this study such as age, education, experience and location of primary residence. A probit model will be used to help determine the impact of these factors on adoption.

Information Available: The variable CLUB represents whether the queen breeder is a member of a local beekeeping club. Involvement with sources of knowledge such as clubs and related

organizations are considered to significantly affect adoption. Many studies have shown that improved information helps facilitate adoption. These include farmer associations (Caviglia and

24

Kahn, 2001), organizations (Arellanes and Lee, 2003), and information gathered by other farmers (Foster and Rosenzweig, 1995). It is hypothesized that participation in beekeeping clubs on a regular basis has a positive relationship on the likelihood of VSH adoption.

Risk Preference: Risk, RISK, is a key indicator to include in the model because it serves to determine how the risk preference of the potential adopter impacts adoption. Respondents were asked how they perceive risk and how they potentially behave with investment decisions. Based on previous literature on risk and uncertainty, risk aversion is hypothesized to be negatively associated with technology adoption (Marra and Carlson, 2002; Hardaker et al., 2004). Risk preference may impact VSH adoption depending on the investment decisions of the potential adopter. In the survey, respondents were asked, “Relative to other investors, how would you characterize yourself?”, (Fausti and Gillespie, 2000). Options consisted of risk taking, risk neutral or risk averse. Depending on the specific characterization of the adopter, risk preference may influence adoption behavior.

Farm Size: One variable represents farm attributes: number of colonies kept in 2011, COLONY. Since the cost of acquiring technology information for a large farm is similar to that of a small farm, there will be a lower cost per unit of area on the larger farm (Perrin et al., 1976). From this, I

hypothesize that queen breeders with higher numbers of colonies may be able to disperse cost across their operation and are expected to more likely adopt VSH bees. Farm size is usually included in studies of adoption evaluation since larger farms may have the advantage of having access to more information sources (Marra and Carlson, 2002). Because of this association, if the VSH queen producer sold to commercial farms as opposed to smaller, hobbyist farms, I hypothesize they are more likely to adopt VSH technology.

Farmer Characteristics: Four variables represent farmer demographics: experience of breeding and selling queens commercially, EXPER; age, AGE (in years); the level of education of the breeder,

25

EDUC, and the location of residence of the queen breeder, SOUTH. Evidence points to the influence of age in the adoption process (Harrison and Ranier, 1992). Hammet et al. (1992) found age had a negative effect on lumber mill export participation, while Ervin and Ervin (1982) found age had a positive association with soil conservation adoption practices. There are some linkages between age and experience in previous studies (Nagubadi, et al., 1996; Agarwal, et al., 1999), including a study whose results suggest age of the individual or length of tenure in the workforce has a negative

Table 3.1. Description of the variables and definitions used in the analysis.

Variable Description

Dependent

VSHX 1 if respondent adopted VSH queen bees in 2012; 0 if otherwise

Independent

Information Sources

CLUB 1 if respondent is a member of a local club or organization; 0 if otherwise

Risk Preference

RISK Relative to other investors, how would you characterize yourself? (Fausti

and Gillespie, 2000). 1 if respondent characterizes themselves as risk averse; 0 if otherwise

Farm Size

COLONY Number of bee colonies respondent kept in 2011

Demographic Variables

EXPER 1 if the years of experience of breeding or selling queens commercially was

greater than 3 years; 0 if less than or equal to 3 years

SOUTH 1 if respondent’s state of primary residence is located in the southern states:

MD, DE, DC, WV, VA, NC, SC, WV, KY, GA, AL, MS, FL, LA, AR, OK, TX; 0 if otherwise

AGE Respondent’s age in years

EDUC 1 if respondent holds a bachelor’s degree or higher; 0 if respondent has some

college, technical school or less Income

INCOME 1 if respondent’s household income was less than $30,000 in 2011,

2 if $30,000 to $59,999, 3 if $60,000 to $89,000, 4 if $90,000 to $119,000, 5 if $120,000 or greater

26

association and/or are more susceptible to negative interference under changing conditions of technology innovation (Agarwal, et al., 1999). I hypothesize that age does not have a significant relationship due to the fact that mostly any person can start breeding queens at any age of their life. Experience in breeding queens, however, is hypothesized to have in a positive influence in the probability of VSH queen adoption.

In the present study, it is hypothesized that a positive relationship exists between education level and adoption. Agarwal et al.'s (1999) hypothesis of the relationship between education and technology innovation states: "Level of education is positively associated with ease of use and usefulness beliefs about an information technology innovation." The level of education has been shown to be positively associated with innovation in other studies (Ersado et al., 2004; Rogers and Shoemaker, 1971). Rogers (2003) describes a degree of communication by interpersonal channels which involve a face-to-face exchange between two or more individuals. The location variable, SOUTH, will help give more insight of the information of VSH queens travel across regions. It is expected that location of residence will be significantly and positively influenced on the adoption decision having originated in the south and possibly disseminating throughout the US. The states chosen for the southern region were based upon the United States Census Bureau census map.

Income: Utility is a measure of happiness that an individual receives from the consumption of a good or service (Pindyck and Rubinfeld, 2009). A higher income level allows one to consume more of those goods and services. In return, utility may increase with the level of income. Income also helps overcome capitals constraint or finance the purchase of an innovation (Feder, et al., 1985). Kebede et al. (1990) found income had a positive effect on the probability of adoption of single-ox, fertilizer and pesticide technologies in developing countries. It is hypothesized that higher household income will positively influence the probability of VSH technology adoption.

27